U.S. Airline Pilot Retention Analysis

General Factor Rankings and Subfactor Priorities

Authors
Affiliations

Michael J. Hickey

University of North Dakota

Marina Efthymiou

Dublin City University

James Higgins

University of North Dakota

Aman Gupta

Embry-Riddle Aeronautical University

Robert Walton

Embry-Riddle Aeronautical University

Published

November 21, 2025

1 Overview

Repository: https://github.com/mikehickey2/US-Airline-Pilot-Retention-Analysis

This analysis examines pilot retention priorities from a survey of U.S. airline pilots (n = 76). The analysis has two parts:

  1. General Retention Factors - Rankings of six broad retention constructs and demographic comparisons
  2. Subfactor Analysis - Within-construct priorities for 31 specific retention elements
Figure 1: Factors of Pilot Retention in U.S. Airlines

2 Statistical Methods

2.1 Sample and Data Collection

Survey data was collected from U.S. airline pilots in May 2025. After exclusions for incomplete responses and non-air carrier pilots, n = 76 valid responses remained for analysis.

2.2 Statistical Tests

All ranking data is ordinal; therefore, non-parametric tests are used throughout:

  • Friedman Test: Within-subjects comparison of factor rankings
  • Mann-Whitney U: Between-groups comparison for binary demographics
  • Kruskal-Wallis H: Between-groups comparison for multi-level demographics

2.3 Multiple Testing Correction

This quantitative study employs a hierarchical false discovery rate (FDR) control procedure:

General Factor Analysis (30 tests): Benjamini-Hochberg FDR at q = 0.10 applied across all demographic comparisons.

Subfactor Analysis (155 tests): Two-step hierarchical procedure:

  1. Screening: For each construct, compute minimum p-value across demographics. Apply BH-FDR at q = 0.10 to these 6 screening p-values.
  2. Within-construct testing: For constructs passing screening, apply Holm’s procedure at alpha = 0.05 to demographic comparisons.

Rationale: Standard Bonferroni correction (alpha/155 = 0.0003) is overly conservative for exploratory research, causing 50-70% Type II error rates. FDR control at q = 0.10 balances discovery against false positives appropriately for hypothesis generation.

2.4 Effect Sizes

  • Mann-Whitney U: Rank-biserial correlation r (small < .3, medium .3-.5, large > .5)
  • Kruskal-Wallis: Eta-squared (small < .01, medium .01-.06, large > .06)

2.5 Software

Analysis conducted in R 4.5.1 using: tidyverse, rstatix, coin, DescTools, knitr, janitor.

Code Availability: The complete analysis code is available at scripts/analysis/unified_retention_analysis.qmd in the GitHub repository: https://github.com/mikehickey2/US-Airline-Pilot-Retention-Analysis

3 Setup

4 Data Loading and Preprocessing

Data Exclusion Summary
exclusion_reason n
did_not_finish 35
included 76
not_air_carrier 2
Table 1: Sample Demographic Summary
Demographic Category n
Age Group ≤ 35 years 58
> 35 years 18
Experience Q1 (least) 19
Q2 19
Q3 19
Q4 (most) 19
Gender Male 67
Female 8
Other/NA 1
Military Background Yes 7
No 69
Position Captain 26
First Officer 50
Carrier Type Regional 57
Major 11
LCC 5
Other 3

Final sample: 76 pilots included in analysis.

5 Part 1: General Retention Factor Analysis

This section analyzes the rankings of six broad retention constructs:

  1. Financial (salary, benefits, bonuses)
  2. Quality of Life/Lifestyle (schedule, family, work-life balance)
  3. Professional Opportunity (stability, upgrade, aircraft type)
  4. Recognition (uniform, respect, awards)
  5. Schedule (fixed vs variable, bidding systems)
  6. Operational (SOPs, training, equipment, pilot skill)

5.1 Descriptive Statistics

Table 1: Rank Summary for General Retention Factors
factor n median iqr mean sd
financial 76 2 2 2.22 1.10
lifestyle 76 2 2 1.99 1.16
schedule 76 2 1 2.46 1.01
operational 76 4 1 4.41 1.07
professional 76 5 1 4.37 1.16
recognition 76 6 1 5.55 0.84
Table 2: Number (and %) of Pilots Ranking Each Factor #1
factor n percent
lifestyle 34 44.7
financial 25 32.9
schedule 13 17.1
professional 3 3.9
operational 1 1.3

5.2 Friedman Test: Overall Factor Rankings

Friedman Test: Overall Difference in Factor Rankings
.y. n statistic df p method
rank 76 234.316 5 0 Friedman test
Table 3: Pairwise Wilcoxon Tests Between General Factors (Holm-adjusted)
comparison statistic p p.adj sig
financial vs lifestyle 1700.5 0.209 0.426
financial vs operational 206.5 0.000 0.000 *
financial vs professional 173.0 0.000 0.000 *
financial vs recognition 10.5 0.000 0.000 *
financial vs schedule 1187.0 0.142 0.426
lifestyle vs operational 104.0 0.000 0.000 *
lifestyle vs professional 149.5 0.000 0.000 *
lifestyle vs recognition 45.0 0.000 0.000 *
lifestyle vs schedule 1009.0 0.016 0.062
operational vs professional 1437.0 0.892 0.892
operational vs recognition 477.0 0.000 0.000 *
operational vs schedule 2793.0 0.000 0.000 *
professional vs recognition 427.0 0.000 0.000 *
professional vs schedule 2667.0 0.000 0.000 *
recognition vs schedule 2905.5 0.000 0.000 *

5.3 Demographic Comparisons

5.3.1 Age Groups (≤35 vs >35)

Table 4: Mann-Whitney U Tests - Age Groups (≤35 vs >35)
factor > 35_n ≤ 35_n > 35_median ≤ 35_median statistic p_raw p_adj_fdr sig_fdr10 n1 n2
lifestyle 18 58 2.5 1 726.0 0.008 0.049 * 18 58
financial 18 58 2.0 2 436.0 0.276 0.736 18 58
schedule 18 58 2.0 2 466.0 0.477 0.736 18 58
professional 18 58 4.0 5 479.0 0.584 0.736 18 58
recognition 18 58 6.0 6 488.5 0.613 0.736 18 58
operational 18 58 4.0 4 505.0 0.832 0.832 18 58

5.3.2 Experience Quartiles

Table 5: Median Rank of General Factors by Experience Quartile
factor Q1 (least)_n Q2_n Q3_n Q4 (most)_n Q1 (least)_median Q2_median Q3_median Q4 (most)_median
financial 19 19 19 19 2 2 2 2
lifestyle 19 19 19 19 1 2 2 2
operational 19 19 19 19 5 4 4 4
professional 19 19 19 19 4 5 5 5
recognition 19 19 19 19 6 6 6 6
schedule 19 19 19 19 3 2 2 2
Table 6: Kruskal-Wallis Tests for Experience Quartiles (FDR-adjusted)
factor statistic df p_raw p_adj_fdr sig_fdr10
financial 0.551 3 0.908 0.908
lifestyle 3.854 3 0.278 0.834
operational 1.364 3 0.714 0.908
professional 1.015 3 0.798 0.908
recognition 4.639 3 0.200 0.834
schedule 2.619 3 0.454 0.908

5.3.3 Gender (Male vs Female)

Table 7: Median Rank of General Factors by Gender
factor Female_n Male_n Female_median Male_median
financial 8 67 3 2
lifestyle 8 67 1 2
operational 8 67 4 5
professional 8 67 5 5
recognition 8 67 6 6
schedule 8 67 2 2
Table 8: Mann-Whitney U Tests - Male vs Female (Exact test, FDR-adjusted)
factor statistic p_raw p_adj_fdr sig_fdr10 n1 n2
financial 365 0.084 0.168 8 67
lifestyle 212 0.312 0.353 8 67
operational 157 0.047 0.168 8 67
professional 320 0.353 0.353 8 67
recognition 356 0.061 0.168 8 67
schedule 192 0.174 0.261 8 67
Note

Note on Gender Comparisons: With only 8 female respondents, statistical power for detecting gender differences is limited. Results should be interpreted cautiously.

5.3.4 Military Background

Table 9: Median Rank of General Factors by Military Background
factor No_n Yes_n No_median Yes_median
financial 69 7 2 2
lifestyle 69 7 2 2
operational 69 7 4 5
professional 69 7 5 4
recognition 69 7 6 6
schedule 69 7 2 2
Table 10: Mann-Whitney U Tests - Military vs Civilian (FDR-adjusted)
factor statistic p_raw p_adj_fdr sig_fdr10 n1 n2
financial 220.5 0.701 0.841 69 7
lifestyle 251.5 0.856 0.856 69 7
operational 211.5 0.578 0.841 69 7
professional 277.5 0.502 0.841 69 7
recognition 198.5 0.338 0.841 69 7
schedule 284.5 0.423 0.841 69 7

5.3.5 Flight-Deck Position (Captain vs First Officer)

Table 11: Median Rank of General Factors by Flight-Deck Position
factor Captain_n First Officer_n Captain_median First Officer_median
financial 26 50 2 2
lifestyle 26 50 2 2
operational 26 50 5 4
professional 26 50 4 5
recognition 26 50 6 6
schedule 26 50 2 2
Table 12: Mann-Whitney U Tests - Captain vs First Officer (FDR-adjusted)
factor statistic p_raw p_adj_fdr sig_fdr10 n1 n2
financial 489.0 0.067 0.200 26 50
lifestyle 723.0 0.398 0.478 26 50
operational 865.0 0.014 0.082 * 26 50
professional 543.0 0.219 0.438 26 50
recognition 573.5 0.296 0.444 26 50
schedule 669.0 0.832 0.832 26 50

5.3.6 Carrier Type

Table 13: Median Rank of General Factors by Carrier Type
factor LCC_n Major_n Other_n Regional_n LCC_median Major_median Other_median Regional_median
financial 5 11 3 57 3 2 1 2
lifestyle 5 11 3 57 2 1 2 2
operational 5 11 3 57 4 4 4 4
professional 5 11 3 57 5 5 4 5
recognition 5 11 3 57 6 6 6 6
schedule 5 11 3 57 1 3 3 2
Table 14: Kruskal-Wallis Tests Across Carrier Types (FDR-adjusted)
factor statistic df p_raw p_adj_fdr sig_fdr10
financial 5.451 3 0.142 0.426
lifestyle 1.479 3 0.687 0.687
operational 2.792 3 0.425 0.661
professional 2.304 3 0.512 0.661
recognition 2.102 3 0.551 0.661
schedule 9.233 3 0.026 0.158

5.4 Summary: General Factor Results

Summary: General Factor Demographic Comparisons (BH-FDR at q=0.10)
Demographic Tests Min_p_raw Min_p_adj Significant_q10
Age (≤35 vs >35) 6 0.0081 0.0488 1
Experience Quartiles 6 0.2000 0.8340 0
Gender (Male vs Female) 6 0.0466 0.1680 0
Military Background 6 0.3380 0.8412 0
Flight-Deck Position 6 0.0136 0.0816 1
Carrier Type 6 0.0264 0.1584 0

General Factor Analysis Summary:

  • Total demographic comparisons: 36
  • Significant at FDR q=0.10: 2
  • Discovery rate: 5.6%

6 Part 2: Subfactor Analysis

This section examines the 31 subfactors within each retention construct and tests for demographic differences in within-construct priorities.

6.1 Financial Subfactors

Financial Subfactors - Descriptive Statistics
item n median mean sd pct_rank1
Competitive salary 76 2.0 1.64 0.74 47.37
Allowances and soft pay 76 4.0 3.96 1.24 1.32
Benefits package 76 3.0 3.36 1.14 5.26
Disability insurance 76 4.5 4.41 1.05 1.32
Job security 76 2.0 2.09 1.28 44.74
New hire/longevity bonuses 76 6.0 5.54 0.92 0.00
Financial: Friedman Test
.y. n statistic df p method
rank 76 231.06 5 0 Friedman test

Financial: Age Group Comparisons
item statistic p p_adj_holm p.adj.signif effsize magnitude
Competitive salary 460.5 0.406 1.000 ns 0.096 small
Allowances and soft pay 502.0 0.806 1.000 ns 0.029 small
Benefits package 583.0 0.444 1.000 ns 0.088 small
Disability insurance 515.0 0.934 1.000 ns 0.010 small
Job security 516.5 0.948 1.000 ns 0.008 small
New hire/longevity bonuses 619.0 0.127 0.762 ns 0.176 small
Financial: Gender Comparisons
item statistic p p_adj_holm p.adj.signif effsize magnitude
Competitive salary 363.0 0.069 0.415 ns 0.211 small
Allowances and soft pay 332.5 0.258 1.000 ns 0.132 small
Benefits package 204.0 0.259 1.000 ns 0.131 small
Disability insurance 174.0 0.093 0.464 ns 0.195 small
Job security 205.0 0.255 1.000 ns 0.133 small
New hire/longevity bonuses 311.0 0.339 1.000 ns 0.112 small
Financial: Position Comparisons
item statistic p p_adj_holm p.adj.signif effsize magnitude
Competitive salary 603.5 0.574 1.000 ns 0.065 small
Allowances and soft pay 596.0 0.547 1.000 ns 0.070 small
Benefits package 646.5 0.973 1.000 ns 0.005 small
Disability insurance 730.0 0.362 1.000 ns 0.105 small
Job security 674.5 0.781 1.000 ns 0.033 small
New hire/longevity bonuses 755.5 0.137 0.822 ns 0.171 small
Financial: Military Comparisons
item statistic p p_adj_holm p.adj.signif effsize magnitude
Competitive salary 311.5 0.164 0.984 ns 0.161 small
Allowances and soft pay 233.5 0.890 1.000 ns 0.017 small
Benefits package 264.5 0.676 1.000 ns 0.049 small
Disability insurance 229.0 0.821 1.000 ns 0.027 small
Job security 169.5 0.173 0.984 ns 0.157 small
New hire/longevity bonuses 246.0 0.926 1.000 ns 0.012 small
Financial: Experience Comparisons
item statistic df p p_adj_holm p.adj.signif effsize magnitude
Competitive salary 1.885 3 0.597 1 ns -0.015 small
Allowances and soft pay 3.862 3 0.277 1 ns 0.012 small
Benefits package 3.708 3 0.295 1 ns 0.010 small
Disability insurance 4.724 3 0.193 1 ns 0.024 small
Job security 2.263 3 0.520 1 ns -0.010 small
New hire/longevity bonuses 4.884 3 0.181 1 ns 0.026 small

6.2 Quality of Life Subfactors

Quality of Life Subfactors - Descriptive Statistics
item n median mean sd pct_rank1
Predictable schedule 76 2 2.51 1.13 19.74
Vacation time 76 4 4.12 1.18 1.32
Family-friendly policies 76 5 4.53 1.28 1.32
Possibility of being based at home 76 2 2.70 1.87 39.47
Travel benefits 76 5 4.87 1.20 1.32
Work-life balance 76 2 2.28 1.41 36.84
Quality of Life: Friedman Test
.y. n statistic df p method
rank 76 139.489 5 0 Friedman test

Quality of Life: Age Group Comparisons
item statistic p p_adj_holm p.adj.signif effsize magnitude
Predictable schedule 329.5 0.015 0.088 ns 0.281 small
Vacation time 544.5 0.780 1.000 ns 0.033 small
Family-friendly policies 481.0 0.610 1.000 ns 0.059 small
Possibility of being based at home 529.5 0.929 1.000 ns 0.011 small
Travel benefits 570.0 0.543 1.000 ns 0.071 small
Work-life balance 640.0 0.134 0.670 ns 0.173 small
Quality of Life: Gender Comparisons
item statistic p p_adj_holm p.adj.signif effsize magnitude
Predictable schedule 197.5 0.212 1.000 ns 0.145 small
Vacation time 253.5 0.803 1.000 ns 0.030 small
Family-friendly policies 265.5 0.972 1.000 ns 0.005 small
Possibility of being based at home 259.0 0.880 1.000 ns 0.019 small
Travel benefits 359.5 0.102 0.612 ns 0.190 small
Work-life balance 290.5 0.693 1.000 ns 0.047 small
Quality of Life: Position Comparisons
item statistic p p_adj_holm p.adj.signif effsize magnitude
Predictable schedule 722.0 0.415 1 ns 0.094 small
Vacation time 604.0 0.605 1 ns 0.060 small
Family-friendly policies 739.5 0.315 1 ns 0.116 small
Possibility of being based at home 556.0 0.287 1 ns 0.123 small
Travel benefits 576.5 0.402 1 ns 0.097 small
Work-life balance 711.5 0.485 1 ns 0.081 small
Quality of Life: Military Comparisons
item statistic p p_adj_holm p.adj.signif effsize magnitude
Predictable schedule 193.5 0.375 1.000 ns 0.103 small
Vacation time 216.0 0.641 1.000 ns 0.055 small
Family-friendly policies 200.5 0.453 1.000 ns 0.087 small
Possibility of being based at home 357.5 0.031 0.186 ns 0.249 small
Travel benefits 236.5 0.932 1.000 ns 0.011 small
Work-life balance 198.0 0.420 1.000 ns 0.094 small
Quality of Life: Experience Comparisons
item statistic df p p_adj_holm p.adj.signif effsize magnitude
Predictable schedule 7.452 3 0.059 0.353 ns 0.062 moderate
Vacation time 5.189 3 0.158 0.790 ns 0.030 small
Family-friendly policies 4.905 3 0.179 0.790 ns 0.026 small
Possibility of being based at home 2.634 3 0.451 1.000 ns -0.005 small
Travel benefits 1.214 3 0.750 1.000 ns -0.025 small
Work-life balance 1.712 3 0.634 1.000 ns -0.018 small

6.3 Professional Opportunity Subfactors

Professional Subfactors - Descriptive Statistics
item n median mean sd pct_rank1
Financially stable airline 76 1 1.18 0.42 82.89
Opportunity for rapid upgrade 76 3 3.45 1.09 1.32
Opportunity to fly larger aircraft 76 3 3.26 1.43 10.53
Promotion/upgrade based on merit 76 3 3.45 1.24 5.26
Upgrade based on length of service 76 4 3.66 0.97 0.00
Professional: Friedman Test
.y. n statistic df p method
rank 76 127.663 4 0 Friedman test

Professional: Age Group Comparisons
item statistic p p_adj_holm p.adj.signif effsize magnitude
Financially stable airline 442.5 0.139 0.695 ns 0.171 small
Opportunity for rapid upgrade 560.0 0.636 1.000 ns 0.055 small
Opportunity to fly larger aircraft 558.5 0.650 1.000 ns 0.053 small
Promotion/upgrade based on merit 495.0 0.739 1.000 ns 0.039 small
Upgrade based on length of service 498.5 0.769 1.000 ns 0.034 small
Professional: Gender Comparisons
item statistic p p_adj_holm p.adj.signif effsize magnitude
Financially stable airline 253.0 0.705 1.000 ns 0.045 small
Opportunity for rapid upgrade 316.5 0.394 1.000 ns 0.099 small
Opportunity to fly larger aircraft 355.5 0.123 0.492 ns 0.179 small
Promotion/upgrade based on merit 173.5 0.097 0.484 ns 0.193 small
Upgrade based on length of service 216.5 0.360 1.000 ns 0.107 small
Professional: Position Comparisons
item statistic p p_adj_holm p.adj.signif effsize magnitude
Financially stable airline 593.5 0.348 1 ns 0.109 small
Opportunity for rapid upgrade 696.5 0.603 1 ns 0.060 small
Opportunity to fly larger aircraft 679.5 0.743 1 ns 0.038 small
Promotion/upgrade based on merit 645.0 0.960 1 ns 0.006 small
Upgrade based on length of service 578.0 0.414 1 ns 0.094 small
Professional: Military Comparisons
item statistic p p_adj_holm p.adj.signif effsize magnitude
Financially stable airline 287.0 0.216 0.648 ns 0.144 small
Opportunity for rapid upgrade 323.5 0.130 0.520 ns 0.175 small
Opportunity to fly larger aircraft 135.0 0.049 0.246 ns 0.227 small
Promotion/upgrade based on merit 229.0 0.824 0.824 ns 0.027 small
Upgrade based on length of service 303.0 0.253 0.648 ns 0.132 small
Professional: Experience Comparisons
item statistic df p p_adj_holm p.adj.signif effsize magnitude
Financially stable airline 1.667 3 0.644 1.000 ns -0.019 small
Opportunity for rapid upgrade 2.979 3 0.395 1.000 ns 0.000 small
Opportunity to fly larger aircraft 0.302 3 0.960 1.000 ns -0.037 small
Promotion/upgrade based on merit 2.428 3 0.488 1.000 ns -0.008 small
Upgrade based on length of service 6.135 3 0.105 0.525 ns 0.044 small

6.4 Recognition Subfactors

Recognition Subfactors - Descriptive Statistics
item n median mean sd pct_rank1
Professional-looking uniforms 76 3 2.80 0.97 14.47
Additional vacation for long service 76 2 2.04 1.01 36.84
Recognition as a professional 76 2 1.78 0.84 44.74
Airline recognition 76 4 3.38 0.88 3.95
Recognition: Friedman Test
.y. n statistic df p method
rank 76 73.168 3 0 Friedman test

Recognition: Age Group Comparisons
item statistic p p_adj_holm p.adj.signif effsize magnitude
Professional-looking uniforms 577 0.476 0.900 ns 0.082 small
Additional vacation for long service 345 0.023 0.093 ns 0.261 small
Recognition as a professional 601 0.300 0.900 ns 0.120 small
Airline recognition 579 0.430 0.900 ns 0.091 small
Recognition: Gender Comparisons
item statistic p p_adj_holm p.adj.signif effsize magnitude
Professional-looking uniforms 183.0 0.120 0.360 ns 0.180 small
Additional vacation for long service 247.5 0.718 1.000 ns 0.043 small
Recognition as a professional 264.0 0.948 1.000 ns 0.009 small
Airline recognition 384.0 0.023 0.091 ns 0.264 small
Recognition: Position Comparisons
item statistic p p_adj_holm p.adj.signif effsize magnitude
Professional-looking uniforms 767.5 0.171 0.513 ns 0.158 small
Additional vacation for long service 496.5 0.078 0.312 ns 0.203 small
Recognition as a professional 724.0 0.385 0.770 ns 0.100 small
Airline recognition 618.0 0.694 0.770 ns 0.046 small
Recognition: Military Comparisons
item statistic p p_adj_holm p.adj.signif effsize magnitude
Professional-looking uniforms 119.5 0.020 0.078 ns 0.269 small
Additional vacation for long service 355.5 0.032 0.096 ns 0.247 small
Recognition as a professional 214.0 0.600 1.000 ns 0.061 small
Airline recognition 274.0 0.511 1.000 ns 0.077 small
Recognition: Experience Comparisons
item statistic df p p_adj_holm p.adj.signif effsize magnitude
Professional-looking uniforms 4.887 3 0.180 0.540 ns 0.026 small
Additional vacation for long service 8.274 3 0.041 0.163 ns 0.073 moderate
Recognition as a professional 1.184 3 0.757 1.000 ns -0.025 small
Airline recognition 0.637 3 0.888 1.000 ns -0.033 small

6.5 Schedule Subfactors

Schedule Subfactors - Descriptive Statistics
item n median mean sd pct_rank1
Fixed schedule 76 3.5 3.18 1.64 25.00
Variable schedule 76 3.0 3.39 1.39 14.47
Flexible work rules 76 2.0 2.12 1.31 42.11
Bid line seniority system 76 3.0 2.82 1.27 18.42
Vacation bidding rules 76 3.0 3.49 0.97 0.00
Schedule: Friedman Test
.y. n statistic df p method
rank 76 37.632 4 0 Friedman test

Schedule: Age Group Comparisons
item statistic p p_adj_holm p.adj.signif effsize magnitude
Fixed schedule 464.5 0.472 1.00 ns 0.083 small
Variable schedule 493.5 0.724 1.00 ns 0.041 small
Flexible work rules 669.0 0.058 0.29 ns 0.218 small
Bid line seniority system 574.5 0.514 1.00 ns 0.076 small
Vacation bidding rules 479.0 0.588 1.00 ns 0.063 small
Schedule: Gender Comparisons
item statistic p p_adj_holm p.adj.signif effsize magnitude
Fixed schedule 202.5 0.248 0.992 ns 0.134 small
Variable schedule 387.0 0.036 0.180 ns 0.243 small
Flexible work rules 229.5 0.489 0.992 ns 0.081 small
Bid line seniority system 270.5 0.972 0.992 ns 0.005 small
Vacation bidding rules 204.0 0.256 0.992 ns 0.132 small
Schedule: Position Comparisons
item statistic p p_adj_holm p.adj.signif effsize magnitude
Fixed schedule 606.0 0.622 1 ns 0.057 small
Variable schedule 647.0 0.977 1 ns 0.004 small
Flexible work rules 714.0 0.462 1 ns 0.085 small
Bid line seniority system 747.5 0.276 1 ns 0.126 small
Vacation bidding rules 551.5 0.263 1 ns 0.129 small
Schedule: Military Comparisons
item statistic p p_adj_holm p.adj.signif effsize magnitude
Fixed schedule 177.0 0.235 1 ns 0.137 small
Variable schedule 286.5 0.410 1 ns 0.096 small
Flexible work rules 227.5 0.797 1 ns 0.031 small
Bid line seniority system 216.0 0.645 1 ns 0.054 small
Vacation bidding rules 257.0 0.779 1 ns 0.033 small
Schedule: Experience Comparisons
item statistic df p p_adj_holm p.adj.signif effsize magnitude
Fixed schedule 1.454 3 0.693 1.000 ns -0.021 small
Variable schedule 0.483 3 0.923 1.000 ns -0.035 small
Flexible work rules 4.079 3 0.253 0.968 ns 0.015 small
Bid line seniority system 4.188 3 0.242 0.968 ns 0.016 small
Vacation bidding rules 5.826 3 0.120 0.600 ns 0.039 small

6.6 Operational Subfactors

Operational Subfactors - Descriptive Statistics
item n median mean sd pct_rank1
Unambiguous SOPs 76 3 3.07 1.59 28.95
Proactive training environment 76 3 2.84 1.17 11.84
Well-maintained aircraft 76 2 2.25 1.16 35.53
Well-equipped aircraft 76 5 4.13 1.18 1.32
Highly skilled fellow pilots 76 3 2.71 1.24 22.37
Operational: Friedman Test
.y. n statistic df p method
rank 76 59.463 4 0 Friedman test

Operational: Age Group Comparisons
item statistic p p_adj_holm p.adj.signif effsize magnitude
Unambiguous SOPs 486.0 0.655 1.000 ns 0.052 small
Proactive training environment 406.0 0.145 0.725 ns 0.168 small
Well-maintained aircraft 556.0 0.671 1.000 ns 0.050 small
Well-equipped aircraft 551.5 0.692 1.000 ns 0.046 small
Highly skilled fellow pilots 608.0 0.283 1.000 ns 0.124 small
Operational: Gender Comparisons
item statistic p p_adj_holm p.adj.signif effsize magnitude
Unambiguous SOPs 255.0 0.825 1 ns 0.027 small
Proactive training environment 240.0 0.626 1 ns 0.057 small
Well-maintained aircraft 237.0 0.586 1 ns 0.064 small
Well-equipped aircraft 328.5 0.247 1 ns 0.135 small
Highly skilled fellow pilots 290.0 0.705 1 ns 0.045 small
Operational: Position Comparisons
item statistic p p_adj_holm p.adj.signif effsize magnitude
Unambiguous SOPs 620.0 0.740 1 ns 0.039 small
Proactive training environment 722.5 0.416 1 ns 0.094 small
Well-maintained aircraft 618.0 0.720 1 ns 0.042 small
Well-equipped aircraft 637.5 0.883 1 ns 0.018 small
Highly skilled fellow pilots 655.5 0.955 1 ns 0.007 small
Operational: Military Comparisons
item statistic p p_adj_holm p.adj.signif effsize magnitude
Unambiguous SOPs 287.0 0.405 1 ns 0.096 small
Proactive training environment 280.0 0.481 1 ns 0.082 small
Well-maintained aircraft 218.0 0.668 1 ns 0.050 small
Well-equipped aircraft 204.5 0.463 1 ns 0.085 small
Highly skilled fellow pilots 225.0 0.768 1 ns 0.035 small
Operational: Experience Comparisons
item statistic df p p_adj_holm p.adj.signif effsize magnitude
Unambiguous SOPs 4.274 3 0.233 0.932 ns 0.018 small
Proactive training environment 0.476 3 0.924 1.000 ns -0.035 small
Well-maintained aircraft 1.985 3 0.575 1.000 ns -0.014 small
Well-equipped aircraft 0.326 3 0.955 1.000 ns -0.037 small
Highly skilled fellow pilots 5.161 3 0.160 0.800 ns 0.030 small

6.7 Hierarchical FDR Screening Results

Construct-Level Screening: Benjamini-Hochberg FDR (q=0.10)
construct n_subfactors min_p rank p_adj_fdr passes_screen decision
Quality of Life 6 0.0147 1 0.0588 TRUE Proceed to detailed testing
Recognition 4 0.0196 2 0.0588 TRUE Proceed to detailed testing
Schedule 5 0.0359 3 0.0718 TRUE Proceed to detailed testing
Professional 5 0.0493 4 0.0739 TRUE Proceed to detailed testing
Financial 6 0.0691 5 0.0829 TRUE Proceed to detailed testing
Operational 5 0.1450 6 0.1450 FALSE No demographic differences detected

Screening Summary:

  • Constructs passing FDR screen (p_adj < 0.10): 5 of 6
Note

For constructs passing screening, demographic comparisons have been adjusted using Holm’s procedure. Tests with p_adj_holm < 0.05 indicate significant demographic differences in subfactor prioritization.

7 Combined Results Summary

Combined Analysis Summary: Multiple Testing Approach
Analysis Total_Tests FDR_Method Significant_FDR Note
General Factors 36 BH q=0.10 2 6 factors × 5 demographics (no Position analysis)
Subfactor Analysis 155 Hierarchical: BH screening + Holm within 5 5 of 6 constructs pass screening

7.1 Summary

General Retention Factors (Part 1):

  • Total demographic comparisons: 36
  • Significant at FDR q=0.10: 2

Subfactor Analysis (Part 2):

  • Constructs analyzed: 6
  • Total subfactors: 31
  • Constructs passing FDR screening: 5
  • Constructs with demographic differences: Quality of Life, Recognition, Schedule, Professional, Financial
TipInterpretation

The hierarchical FDR procedure controls for the 155 subfactor tests by:

  1. Screening at construct level (BH-FDR q=0.10 on 6 tests)
  2. Applying Holm correction within constructs that pass screening

This approach maintains family-wise error rate control while preserving power to detect genuine demographic differences.

8 Limitations

WarningStatistical Power Considerations

With n = 76, this study has limited statistical power. Non-significant results should not be interpreted as evidence of no effect. Many analyses may be underpowered to detect small-to-medium effects.

This analysis has several important limitations:

8.1 Sample Size and Power

The total sample of n = 76 provides adequate power only for detecting large effects (Cohen’s d > 0.8 or r > 0.5). For demographic comparisons:

Comparison Group Sizes Detectable Effect (80% power)
Age groups ~40 vs ~36 Medium (r ≈ 0.35)
Experience ~19 per quartile Large (eta² > 0.10)
Gender 67 vs 8 Very large only (r > 0.6)
Military ~35 vs ~40 Medium (r ≈ 0.35)
Position ~45 vs ~30 Medium (r ≈ 0.35)

Gender comparisons warrant particular caution: With only 8 female respondents, the analysis has approximately 20% power to detect medium effects - meaning 80% of true medium-sized gender differences would go undetected.

8.2 Additional Limitations

  1. Multiple Testing Burden: Despite FDR control, the large number of tests (155 in subfactor analysis) increases false discovery risk; the expected number of false discoveries at q=0.10 is approximately 10% of significant findings
  2. Self-Selection Bias: Survey respondents may not represent all U.S. airline pilots; those with strong opinions about retention may be overrepresented
  3. Cross-Sectional Design: Retention priorities may change over career stage; this snapshot cannot capture temporal dynamics
  4. Ranking Methodology: Forced rankings prevent assessment of absolute importance; factors ranked low may still be important in absolute terms

9 References

9.1 Primary Sources

Hickey, M. J. (2025). Rethinking pilot retention in the United States: An analysis of key factors [Doctoral dissertation, University of North Dakota]. ProQuest Dissertations & Theses. https://www.proquest.com/docview/3246414757

Efthymiou, M., Njoya, E. T., Lo, P. L., Papatheodorou, A., & Randall, D. (2020). The factors influencing entry level airline pilot retention: An empirical study of Ryanair. Journal of Air Transport Management, 91, 101997. https://doi.org/10.1016/j.jairtraman.2020.101997

9.2 Statistical Methods

Benjamini, Y., & Hochberg, Y. (1995). Controlling the false discovery rate: A practical and powerful approach to multiple testing. Journal of the Royal Statistical Society: Series B, 57(1), 289-300.

Bender, R., & Lange, S. (1999). Multiple test procedures other than Bonferroni’s deserve wider use. BMJ, 318(7183), 600-601.

Armstrong, R. A. (2014). When to use the Bonferroni correction. Ophthalmic and Physiological Optics, 34(5), 502-508.

Li, J., & Ghosh, J. K. (2014). A two-step hierarchical hypothesis set testing framework. BMC Bioinformatics, 15, 108.